Analysis of numerical methods for the simulation of deformable models

Simulating deformable objects based on physical laws has become the most popular technique for modeling textiles, skin, or volumetric soft objects like human tissue. The physical model leads to an ordinary differential equation. Recently, several approaches to fast algorithms have been proposed.In t...

Full description

Saved in:
Bibliographic Details
Published in:The Visual computer 2003-12, Vol.19 (7-8), p.581-600
Main Authors: Hauth, Michael, Etzmuss, Olaf, Strasser, Wolfgang
Format: Article
Language:English
Subjects:
Algorithms
Differential equations
Formability
Human tissues
Mathematical models
Newton methods
Numerical methods
Stability
Textiles
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Simulating deformable objects based on physical laws has become the most popular technique for modeling textiles, skin, or volumetric soft objects like human tissue. The physical model leads to an ordinary differential equation. Recently, several approaches to fast algorithms have been proposed.In this work, more profound numerical background about numerical stiffness is provided. Stiff equations impose stability restrictions on a numerical integrator. Some one-step and multistep methods with adequate stability properties are presented. For an efficient implementation, the inexact Newton method is discussed. Applications to 2D and 3D elasticity problems show that the discussed methods are faster and give higher-quality solutions than the commonly used linearized Euler method.
ISSN:0178-2789
1432-2315
DOI:10.1007/s00371-003-0206-2